Run Step-3.7-Flash on M5 Ultra

Step-3.7-Flash runs at 69 tok/s (68.5 as recorded) on a 256 GB M5 Ultra, a community run reported by omlx.ai (MLX, 4-bit, 1k context). It needs about 65 GB of unified memory. It is StepFun's 198B mixture-of-experts model (11B active), released under the Apache 2.0 licence.

Speed
community69
tok/s
First Token
community1.132
seconds
RAM Needed
65
GB minimum
Engine
MLX
of this figure
Want to run Step-3.7-Flash faster — or step up to bigger models? Find your Mac →

The figure

This is a community run: a real run, reported by its source and linked below. The index records it as given and never adjusts it.

MetricValue
Tokens per secondcommunity69 tok/s
Provenancecommunity run · omlx.ai
Quantization4-bit
EngineMLX
Time to first token1.132s
Prompt processing (prefill)community1602 tok/s at 8K context
Context1k
Figure dated2026-10-19
Minimum RAM65 GB
Parameters198B MoE (11B active)
Chip bandwidth1228 GB/s

Every figure for Step-3.7-Flash on the M5 Ultra

Each row as the index holds it, with its source. “Mac RAM” is the memory of the Mac a run used, not the model’s requirement (65 GB). The headline figure follows the reference rule: the lower-median measured plain-decode run at 5 bits or fewer (a run whose quant was not stated counts only when no stated-quant run qualifies), else the estimate; a run marked “not used to anchor estimates” is listed here but never becomes the headline.

SpeedQuantEngineMac RAMContextSourceNote
community68.5 tok/s4-bitMLX256 GB1komlx.ai64-core GPU (1.2 TB/s, the same bandwidth as the 80-core part; Apple sells this chip with 256 GB). oMLX v0.7.1rc1, MTP/DFlash off; Step-3.7-Flash-oQ4-mtp (oMLX's plain 4-bit oQ; MTP head not used); 106.3 GB peak. 67.7 tok/s at 4k, 59.4 at 32k.
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Speed by context depth: Step-3.7-Flash on the M5 Ultra

Measured figures only, each from its source’s own session; the index does not estimate long-context or prefill speed, so a missing cell is —. Each @ column is plain decode with that many tokens already in the prompt, from the same session as the row’s short-context figure: decode slows as the KV cache it reads grows. Prefill is prompt processing, the tokens per second a runtime reads a prompt at; with long agent contexts it sets how long you wait before the first token.

RunDecode (row)@4K@8K@16K@32KPrefillTTFT
MLX · 4-bit · 256 GB Mac · omlx.aicommunity68.5 tok/s
1k
community67.7 tok/scommunity66.6 tok/scommunity62.7 tok/scommunity59.4 tok/scommunity1602 tok/s
at 8K
community1.132 s

Setup: run Step-3.7-Flash on a Mac

Ollama — the index has not verified an Ollama tag for Step-3.7-Flash, so it doesn’t print one (a guessed tag fails with “pull model manifest: file does not exist”). Search the Ollama library →

MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:

pip install -U mlx-lm
mlx_lm.generate --model <repo-id> --prompt "Hello"

Downloads: Hugging Face repo → · Hugging Face model card →

The 69 tok/s figure above came from MLX at 4-bit; other runtimes and quants run at different speeds.

Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.

Step-3.7-Flash on other Apple Silicon chips

ChipSpeedQuantEngineBandwidth
M3 Ultracommunity57 tok/s4-bitMLX819 GB/s
M5 Maxcommunity48 tok/soQ2eMLX614 GB/s
M4 Maxestimated26 tok/sQ4_K_M—546 GB/s

For the estimated figures here, the bandwidth formula’s typical error for large MoEs is ±34% (mostly below the measured figure), against 70 measured runs of 18 models (how it is computed).

About Step-3.7-Flash

StepFun's Apache 2.0 198B MoE (11B active, 23 May 2026), image input, 256K context; it supersedes Step-3.5-Flash. Card: SWE-Bench Pro 56.3 (vendor scaffold), Terminal-Bench 2.1 59.5. The oQ2e MLX build is 63 GB and IQ4_XS ~105 GB, so it needs 96–128 GB; upstream llama.cpp has run it since b9473, and community oMLX runs measured ~48–51 tok/s on an M5 Max at 2–3 bits.

Apache 2.0 licence · 198B MoE · 256K context · LLMCheck capability 41/50 · see its leaderboard rank →

Memory by context length

Estimated, not measured: the index’s minimum for Step-3.7-Flash (65 GB — the weights at the quant that fits, plus ~2 GB) plus the KV cache for that many tokens, computed from its config.json. 12 full layers + 33 sliding-window layers (window 512). A 16-bit KV cache is the default in llama.cpp, Ollama, LM Studio and mlx-lm; a q8 cache halves the KV part. Coding agents hold 64K tokens or more. Runtimes also add a buffer that grows with context, so leave a few GB spare. macOS lets the GPU use only part of a Mac’s memory by default (figures for an M5 Ultra Mac, which runs macOS 26 or later): “fits” means within the GPU memory a Mac of that tier can use with the limit raised, a tier tagged raise GPU limit fits only after sudo sysctl iogpu.wired_limit_mb raises it (the tooltip has the exact command), and the last column is the smallest Mac memory size whose default limit holds it (the GPU memory limit).

ContextKV cache (16-bit)RAM neededRAM, q8 KVSmallest tier that fitsWithin the default GPU limit
8K~0.5 GB~65.2 GB~65.1 GB128 GB Mac96 GB Mac
32K~1.7 GB~66.4 GB~65.7 GB128 GB Mac96 GB Mac
64K (coding-agent context)~3.3 GB~68.0 GB~66.5 GB128 GB Mac96 GB Mac
128K~6.5 GB~71.2 GB~68.1 GB128 GB Mac96 GB Mac

System requirements

Macs with the M5 Ultra

🛒 Get a Mac that runs Step-3.7-Flash

Step-3.7-Flash needs about 65 GB of unified memory. These Apple Silicon Macs have enough to run it. Apple builds the higher-memory configurations to order; Amazon does not stock them, so the used option links Amazon Renewed listings.

As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.

Compare More Models

See how Step-3.7-Flash stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard